01
A short call
Twenty minutes, no deck. You describe the system and the problem; I ask the questions that determine whether this is worth either of our time. If it isn't, I say so on that call.
Three ways to work together
Backend and infrastructure only. I do not do design, mobile, or frontend build-out — there are people better at those than me, and pretending otherwise is how projects go wrong.
$1,500 · 5 working days
Go / Rust API performance audit
I profile your service under realistic load, find what is actually slow, and hand back a written report plus a pull request.
Most performance work fails because it starts from a guess. This starts from measurement: pprof or perf against your workload, query plans pulled from your database, allocation profiles where they matter. You get the evidence alongside the conclusion, so your team can check my reasoning rather than take it on faith. If I find nothing worth fixing, the report says exactly that and explains why — the measurements are still yours to keep.
What you get
- Written report: findings ranked by measured impact, not vibes
- Flame graphs and benchmark output from your workload, not a synthetic one
- A pull request implementing the fixes worth making immediately
- Query plans for every statement I touched
- A 60-minute walkthrough call once you have read the report
$6,000 · 4–12 weeks, typically
Backend rewrite or new service build
A production backend built end to end — schema, API, deployment, handover — by the person who writes the code.
The range reflects scope, not seniority tiers. A single well-defined service with a clear data model sits at the lower end; a rewrite that has to run alongside the system it replaces, with dual writes and a migration path, sits higher. I quote after a design document exists, because quoting before that is guesswork billed to you. Work happens in milestones you can stop at.
What you get
- A short design document agreed before implementation starts
- The service itself: typed API surface, migrations, test suite
- CI pipeline and deployment configuration you own outright
- Runbook covering operations, failure modes and rollback
- Handover sessions with whoever maintains it next
$4,000 · Monthly, 30 days notice
Fractional backend lead
Ongoing senior backend judgement for a team that needs it without a full-time hire.
This is for teams who have backend decisions to make and nobody senior enough to make them confidently. You get the judgement without the hiring timeline or the equity conversation. It works well alongside a build engagement — I stay on after delivery so the system does not drift from its design once I stop touching it daily.
What you get
- Architecture review and technical direction for backend work
- Code review on the changes that carry real risk
- Hands-on implementation of the hardest pieces
- Interview support and technical assessment for backend hires
- A standing weekly call plus async availability in your channel
How I work
Four steps, no surprises
The same sequence whether the engagement is five days or three months.
02
A written scope
Before money changes hands you get the scope in writing: what I will do, what I will not, what I need from you, and what the deliverable actually is. For builds this is a design document; for audits it is a one-pager.
03
The work, in the open
You get commits as they happen, not a reveal at the end. Questions get answered in your channel within a working day. I am in IST and overlap comfortably with Europe; US mornings need a little planning.
04
Handover that survives me
A runbook, a walkthrough, and code your team can maintain without me. The measure of the engagement is whether it still works six months after I stop touching it.
Stack
What I run in production
Short on purpose. Everything listed is something I would be comfortable being interrogated about in a technical call.
-
Go
services, 6+ years in production
-
Rust
performance-critical services and tooling
-
TypeScript
frontends and edge functions
-
PostgreSQL
primary datastore, schema and tuning
-
MongoDB
document workloads
-
Redis
caching and queues
-
Valkey
Redis-compatible, open governance
-
Docker
everything ships as an image
-
Kubernetes
orchestration at scale
-
K3s
lightweight self-hosted clusters
-
containerd
the runtime under the orchestrator
-
gRPC
service-to-service contracts
-
Cloudflare
edge, Workers, R2
Not sure which of the three fits?
Describe the problem rather than the solution and I'll tell you which one applies — or that none of them do.
Start a conversation